[OFB Cafe] Xeon Comparions
Timothy Butler
tbutler at ofb.biz
Mon Jul 14 21:03:28 CDT 2008
> I can't point to any hard evidence, but the older chips are likely
> P4/Netburst (not sure if that's even the correct microarchitecture
> codename) and the newer chip is likely a branch of the Core
> architecture.
Yes, that seems to be right on. I did some serious digging and found
that the Xeon 2.8 GHz was a Netburst processor with 400 MHz bus and
2mb L3 cache. The 3040 1.86 GHz is Core-based, has 2mb L2 and a 1066
MHz bus.
> There's also the issue of "peripherals" - memory controller, PCI vs
> PCIe, etc - the newer chip likely has more up-to-date support
> components
> and should reduce some of the bottlenecks the older chips will be
> more
> or less stuck with.
Good points. The older chip also came in a system with PATA vs. SATA.
It's via a datacenter server lease, so I don't have a big worry
component costs, but I settled on the 3040. I originally planned to go
with the 3040, but the 2.8 GHz was on sale, and I thought, "well, for
about the same price, which is the better choice?" Hopefully going
with the newer architecture will make up for being a slower clock.
Based on what you say, which sounds like a familiar calculation, the
1.86 GHz clock should be just about as good, right? It works out
neatly to 2/3 of the clock speed...
Thanks for the advice.
-Tim
---
Timothy R. Butler | "The yellow fog that rubs its back upon the window-
panes,
Editor, OfB.biz | The yellow smoke that rubs its muzzle on the
window-panes
tbutler at ofb.biz | Licked its tongue into the corners of the
evening,
timothybutler.us | Lingered upon the pools that stand in drains."
--
T.S. Eliot
More information about the Cafe
mailing list